25th ARRB Conference – Shaping the future: Linking policy, research and outcomes, Perth, Australia 2012
IMPACT OF CLIMATE RELATED CHANGES IN
TEMPERATURE ON CONCRETE PAVEMENT: A FINITE
ELEMENT STUDY
Dr Gary Chai, Griffith University, Queensland, Australia
Dr Rudi van Staden, Griffith University, Queensland, Australia
Associate Professor Hong Guan, Griffith University, Queensland,
Australia
Professor Yew-Chaye Loo, Griffith University, Queensland, Australia
ABSTRACT
Thermal cracking of concrete is a common cause of deterioration and it is due to temperature
difference constraining concrete contraction. With the addition of repetitive loading, micro-cracks
are expected to propagate through the slab due to both external loadings and erosion in the
cracks. This research aims to develop a three-dimensional finite element model of a jointed
plain concrete pavement system and evaluate stress characteristics under high severity climate
change temperature scenarios for 2007, 2030, 2050 and 2070 in South East Queensland. A
series of static analyses are performed to replicate a standard vehicle axle wheel loads
approaching and leaving the concrete joint. The tensile stress is measured on the top of the
concrete slab at critical locations in both the longitudinal and transverse direction. Results for
thermal-expansive stresses show that the likelihood of cracking in concrete increases
significantly with changes in temperature gradients due to hotter climate causing downward
curling of slab. Further, for temperature loading alone the tensile stress increases by a
maximum of 1.37 MPa from 2007 to 2070.
Keywords: concrete pavement, dowel bar, finite element method, thermal cracking, climate
change.
INTRODUCTION
South East Queensland (SEQ) is Australia’s fastest growing metropolitan region. From 2006 to
2031 its population is expected to grow from 2.8 to 4.4 million (Department of Infrastructure and
Planning 2009). At present the state controlled road network in SEQ is approximately 3,138 km
(Department of Transport and Main Roads 2009) and local governments are responsible for
more than 10,000 km of arterial, sub-arterial and urban roads. SEQ’s population is heavily
urbanised and is generally concentrated in Brisbane, Gold Coast and Sunshine Coast regions.
Based on the Australian trends (CSIRO & BoM 2007), the future climate in SEQ is expected to
become hotter and drier. The region will experience an increase in mean annual temperatures
and decrease in mean annual rainfall.
The topic of climate change has been recognised globally as an issue of utmost concern as the
threat of changes in climate poses problems to all nations of the world (Intergovernmental Panel
on Climate Change 2007). Climate change is expected to pose many challenges to the road
design, construction and maintenance in the region (Serrao-Neumann 2011). The direct impacts
of changes in rainfall patterns can alter moisture balances and influence pavement
deterioration. In addition, temperature changes can affect aging of flexible pavements by
resulting in cracking of the surface, with a consequent loss of waterproofing. The result is that
surface water can enter the flexible pavement causing potholing and fairly rapid loss of surface
condition. Rigid or concrete pavements are impacted by temperature changes through
alterations in expansion and contraction movements and subsequently cracks form.
© ARRB Group Ltd and Authors 2012
1
25th ARRB Conference – Shaping the future: Linking policy, research and outcomes, Perth, Australia 2012
Temperature variations cause curling and thermal-expansion stresses within the concrete.
Curling stresses result from temperature gradients through the slab depth and
thermal-expansion stresses are induced due to uniform changes in temperature that cause the
slab to expand. Indirect impacts of changes in rainfall and temperature include increased
maintenance costs and intervals to increase resilience and subsequent interruptions to traffic.
The load carrying capacity and low maintenance of concrete road means that it is a usual
preference for pavement application subject to high volumes of heavy traffic loads. The concrete
pavement system typically consists of a jointed plain, jointed reinforced or continuously
reinforced concrete layer. The subbase and road base layers usually consist of rock aggregate
which meets certain road specification. This study considers only jointed plain concrete
pavement which uses dowel bars and has joints which are transverse to the direction of travel.
Note that joints are placed in the concrete slab to control cracking and provide space and
freedom for concrete subject to traffic loading and thermal stresses.
Concrete pavements are designed to provide safe and long-lasting road surfaces. However,
fatigue and deterioration (i.e. distress) from repeated vehicular loading, temperature and
moisture changes over time are the most common observed failure mechanisms and major
attributors to pavement maintenance and rehabilitation costs. Thus, a good understanding of
pavement distress is crucial to the successful design and operation of road infrastructure.
Curling deformation, resulting in thermal-expansion stresses in the concrete slab, is a
characteristic phenomenon under environmental and repeated vehicle loads (Huang 2010).
Such deformation may lead to void formation due to the accumulated plastic deformation and
subsequent disengagement of the base course from the concrete. Distortion of the slab due to
both upward and downward curling occurs respectively when the top surface of the slab is
cooler than the base course and also when there is a higher temperature on the top surface
respectively, as illustrated in Figure 1.
Upward
Curling
Downward Curling
Concrete Slab
Base Course
Subgrade
Figure 1: Curling of concrete slab
Distress of the pavement in the form of joint deterioration or cracking also attributes to void
formation by allowing moisture infiltration. The combination of distress and layer voids will
further reduce the pavement load carrying capacity. Friberg (1938) and von Quintus and
Killingsworth (1993) stated that distress is influenced mostly by compressive stress and that the
first sign of deterioration is the formation of transverse or longitudinal cracking within the
concrete slab. It has also been reported that transverse and longitudinal cracking are more
common than D-cracks, corner cracks and meander cracks (Chen et al. 2002; Machemehl
2005). It is frequently understood that the aim of joint design for concrete pavement is to reduce
transverse and longitudinal cracking (CCA 2004).
Cracking of the concrete is an important performance-related factor that needs to be considered
in the concrete pavement design process. Transverse cracks are one of the primary failure
modes and it is due to downward curling and thermal stresses during the day-time. This study
therefore aims to evaluate flexural tensile stress characteristics within the concrete slab to give
further insight into cracking. A finite element model of the symmetrical half of the
three-dimensional (3D) pavement model is developed with realistic vehicular loadings and
restraint conditions. The most critical load conditions will be considered in the finite element
modelling by applying a series of heavy single axle loading and also high severity climate
change related temperature increases for 2030, 2050 and 2070. Ultimately, the findings from
this research will provide information on the detailed behaviour of stresses for different
temperature loadings. This will then help in identifying techniques for reducing cracking by, for
example, shorter joint spacing, thicker slab, stronger flexural strength of concrete, tied shoulders
or wider lanes will aid performance.
© ARRB Group Ltd and Authors 2012
2
25th ARRB Conference – Shaping the future: Linking policy, research and outcomes, Perth, Australia 2012
The cost of construction and maintenance for concrete pavement is high and therefore it is
essential that this road asset is properly monitored during the life of the pavement. From the
early days of concrete pavement analysis, attention has always been concentrated on stresses
induced by wheel loads. However, the impact of changes in climatic conditions in the future,
such as the increase in global temperatures, on the pavement deterioration has not been fully
evaluated. The main objective of this research is to study the performance of the concrete joint
with dowel bars subject to an increase in temperature under the climate change conditions in
South East Queensland. The scope of this study includes:
 developing a 3D finite element model of the symmetrical half of the pavement system
featuring a series of static standard vehicle axle wheel loads approaching and leaving the
concrete joint; and,
 evaluating the thermal-expansion stresses of the concrete slab due to changes in the
temperature gradient based on climate change data in SEQ.
METHODOLOGY
The modelling and simulation herein are performed using Strand7 Finite Element Analysis
(FEA) System (Strand7 2004). The pavement section selected for this study resembles the
typical concrete pavement structure constructed in SEQ. The 3D representation of the three
types of pavement layers, i.e. concrete slab, base course and subgrade, are defined and full
friction is applied at the boundaries. Illustrated in Figure 2 is the finite element model including
vehicular loading, nodal temperatures and restraint conditions applied to the pavement. These
loading and restraint conditions are commonly assumed in previous literature (Al-Hadidy 2009);
Ju 2009).
(a) Finite element model of pavement system
(b) Stress measurement locations
(c) Vehicle load location
Figure 2: Details of symmetrical half of the pavement system model, vehicular loading,
dowel bar and joint
Dowel bars are modelled using beam elements with a defined diameter of 32 mm (see Figure 2
(c)). The mesh size is reduced in the vicinity of the concrete slab-dowel bar interface to aid the
© ARRB Group Ltd and Authors 2012
3
25th ARRB Conference – Shaping the future: Linking policy, research and outcomes, Perth, Australia 2012
accuracy of results. The total numbers of 8 node hexahedral brick elements are 8976 for the
concrete slab, 8624 for the base course and 24640 for the subgrade. The total number of nodal
points for the entire pavement model is 46102. The bottom surface of the pavement is restraint
from deforming in all directions and roller restraints are applied to the sides of the pavement.
The material properties are provided in Table 1 and are assumed to be linear, homogeneous
and elastic in behaviour. The thermal expansion, conductivity and specific heat of the concrete
slab is assumed to be 1×10-5/oC, 1.37×10-3J/s/mm/oC and 880J/kg/oC, respectively.
Table 1: Material properties and layer thicknesses
Description
Concrete
slab
Base
course
Subgrade
Dowel
bar
Plastic sleeve
of dowel bar
28,000
350
50
200,000
16.7
Layer thickness (mm)
250
150
1,900
N/A
N/A
Poisson’s ratio
0.18
0.4
0.4
0.3
0.3
2,400
2,000
1,800
7,830
1,890
Young’s modulus (MPa)
3
Density (kg/m )
Loading conditions
In the single axle dual wheel system shown in Figure 1, each wheel carries a 20 kN load. For
the symmetrical half of the pavement system model, only two 20 kN wheel loads are
considered. The equivalent contact pressure of each of the two tyres is assumed to be
distributed uniformly over a rectangular area of 2.63×10-2m2 (i.e. 759.73 kPa). A total of five
loading cases are considered to replicate a vehicle approaching and leaving the joint as shown
in Figure 3. In the first (LC1), second (LC2), third (LC3), and forth (LC4) load cases, the
respective resultant force is 292.5, 195, 97.5 and 48.75 mm away from the centre of the joint,
and the resultant force of the fifth load case (LC5) lies on the centre of the joint. Note that only
LC1 and LC5 are discussed herein.
Figure 3: Loading scenarios
© ARRB Group Ltd and Authors 2012
4
25th ARRB Conference – Shaping the future: Linking policy, research and outcomes, Perth, Australia 2012
Temperature
The minimum and maximum mean daily temperatures for Burleigh Heads weather station
(station identification 040034) was collected for every month of the years from 1960 to 2099
using the Climate Tool and database provided by Austroads (Austroads 2008). The database
features a wide range of historical climate data from 1960 to 2007 obtained from the Bureau of
Meteorology. The tool provides future climate data from 2008 to 2099 based on the simulations
by CSIRO in 2004.
Since this study is concerned about the possibility of thermal cracking historical temperature
information for February is used because it indicates on average the highest temperatures when
compared to other months. Climate change is expected to bring warmer nights and days to
Queensland (Intergovernmental Panel on Climate Change 2007). This raises the question as to
what thermal-expansion stresses the pavement will be subject to in the future (i.e. 2008 to 2099)
since the stress condition depends on the magnitude of temperature increase and if this
increase occurs during night and/or day. As the first step, this research only considers a
scenario of ‘hotter days’ where future increases in temperature only occur during the daytime
and the night time temperatures would remain relatively unchange. To model the ‘hotter days’
scenario, the entire pavement system is assumed to be at an absolute minimum temperature
(i.e. 20.5 °C for night time) before the maximum (or daytime) temperature is applied to the top
surface of the concrete slab. This will result in downward curling of the pavement. Figure 4
provides the historical and projected temperatures for high, average and lower severity levels of
both maximum and minimum daily temperature in February.
Figure 4: Mean minimum and maximum daily temperatures for Burleigh Heads in February
Figure 4 and Table 2 provides details of the nodal temperatures applied to the top surface of the
concrete slab and the projected increase in maximum temperature under high severity climate
change projections. These temperatures result in thermal-expansion stresses which are
proportional to the difference between the element’s temperature at the top surface and the
night time temperature. For example, in 2030 the thermal-expansion stresses are relative to the
temperature difference between 29.5 and 20.5 °C.
© ARRB Group Ltd and Authors 2012
5
25th ARRB Conference – Shaping the future: Linking policy, research and outcomes, Perth, Australia 2012
Table 2: Historic and climate change projected temperatures
Year
Climate change
projected increase
in maximum
temperature (°C)
Day temperature of
concrete slab
surface (°C)
Mean
Extreme
2007
0.00
26.30
28.80
2030
0.70
-
29.50
2050
1.30
-
30.10
2070
2.00
-
30.80
Night
temperature
(°C)
Difference
between day and
night
temperatures (°C)
5.80 (for mean)
20.50
9.00
9.60
10.30
- Information not provided.
Results
The tensile stresses measured along longitudinal (LD) and transverse (TD) lines are detailed
herein for the historical and future temperature data of February under higher severity climate
change projections and vehicular loading (see Figure 2 (b)). The start and finish points of LD
(i.e. LD1 and LD2) and TD (i.e. TD1 and TD2) are also identified in Figures 2 (b) and 8 for ease
of discussion. Illustrated in Figure 5 are the temperature contours under downward curling as
obtained by steady state heat analyses using Strand7 (2004). Note that the temperature is
assumed to be linearly distributed throughout the concrete slab.
Figure 5: Temperature contour for downward curling
Figures 6 and 7 respectively show the tensile stress distributions along the lines LD and TD.
Associated stress contours are presented in Figures 8 and 9. From Figures 6 and 8 it can be
seen that the stress along line LD increases as the vehicular load approaches the joint from LC1
to LC5. This increase is less obvious on the leave side (i.e. or after a longitudinal distance of
1200 mm along LD). The stress is reduced for LC1 on the approach side because the vehicle
load induces compression on the top concrete surface. Note that the linear homogenous
material properties of the pavement layers means that the stress increases in linear increments
for all the years from 2007 to 2070. The impact of only changing the temperature and without
vehicle loading is shown in Figures 6 and 9. On an average the tensile stress is 1.14 MPa along
LD higher in 2070 than compared to 2007. Importantly, the stress shows an increase just after
900 and 1500 mm along the line LD. This peak could be due to the influence of the dowel bars
and joint.
© ARRB Group Ltd and Authors 2012
6
25th ARRB Conference – Shaping the future: Linking policy, research and outcomes, Perth, Australia 2012
Figure 6: Tensile stress distribution along longitudinal direction
The distribution and contours of tensile stress along the line TD is illustrated in Figures 7 and 8.
As in Figure 6, it can again be seen that the stress increases as the vehicular load approaches
the joint from LC1 to LC5. The distribution of stress is significantly different for LC1 compared to
LC5 mainly due to the vehicle load being positioned on the line TD for LC1. The sudden
decrease at around 750 and 1050 mm is caused by the compressive vehicle loading. The
impact of only changing the temperature and without vehicle loading again shows an average
tensile stress increase of 1.37 MPa along TD from 2007 to 2070.
© ARRB Group Ltd and Authors 2012
7
25th ARRB Conference – Shaping the future: Linking policy, research and outcomes, Perth, Australia 2012
Figure 7: Tensile stress distribution along transverse direction
© ARRB Group Ltd and Authors 2012
8
25th ARRB Conference – Shaping the future: Linking policy, research and outcomes, Perth, Australia 2012
(a) LC1
(b) LC5
Figure 8: Tensile stress contours in concrete slab
In an attempt to show the impact of the change in temperature of 4.5 °C (i.e. 30.8 – 26.3 °C)
from 2007 to 2070, Figure 9 shows the tensile stress and displacement characteristics for LC1.
(a) Countour for 2007
(b) Countour for 2070
Figure 9: Tensile stress contours under no vehicular loading
© ARRB Group Ltd and Authors 2012
9
25th ARRB Conference – Shaping the future: Linking policy, research and outcomes, Perth, Australia 2012
Tables 3 (a) and (b) presents the maximum stresses and associated locations (points A to D in
Figure 9 (a)) along the lines LD and TD. The location of the maximum stress is indicated in both
Table 3 and Figure 8. The failure load is taken as 0.4×√fcu, where fcu = 40.00 MPa and therefore
the maximum tensile stress is 2.53 MPa. The allowable tensile stress is exceeded by all
maximum stresses under climate change condition in year 2030, 2050 and 2070.
Table 3: Maximum tensile stress at top of concrete slab
Year
Increase in maximum
temperature (°C)
Line
Location along
line (mm)
Maximum tensile
stress (MPa)
(a) Load case, LC1
2007
0.00
2030
0.70
2050
1.30
2070
2.00
LD
A
1.87
TD
B
1.74
LD
A
2.85
TD
B
2.62
LD
A
3.03
TD
B
2.79
LD
A
3.25
TD
B
2.99
LD
A
1.90
TD
C
1.90
LD
A
2.88
TD
C
2.86
LD
A
3.06
TD
C
3.04
(b) Load case, LC5
2007
0.00
2030
0.70
2050
1.30
CONCLUSION
A plain jointed concrete pavement with concrete, base course and subgrade layers was
modelled as a 3D finite element model. The impact of changes in temperature for high severity
climate change projections was evaluated under vehicle loading and with the escalation of daily
maximum temperature. Tensile stress was measured along longitudinal and transverse lines on
the top surface of the concrete for all loadings. Compared to the maximum stress magnitude
measured in 2030 an average increase of 0.20 and 0.18 MPa is found for 2050 and 2070
respectively, along the lines LD and TD. Further, only changing the temperature and without
vehicle loading gives an average tensile stress increase of 1.14 MPa along LD and 1.37 MPa
along TD from 2007 to 2070. All maximum stresses for 2030 to 2070 measured along LD and
TD exceeds the failure load defined by 0.4×√fcu. Overall, the impacts of thermal-expansive
stresses are severe under climate change projections and future research should aim to
validate the findings from this theoretical research. Further research may also evaluate the
upward curling and non-linear temperature gradients through the concrete slab.
REFERENCES
Al-Hadidy, A. and Tan, Y.Q. (2009), Mechanistic analysis of ST and SBS-modified flexible
pavements, Construction and Building Materials, 23, 2941-2950.
Austroads (ed.) (2008), Guide to pavement technology, Sydney: Austroads.
© ARRB Group Ltd and Authors 2012
10
25th ARRB Conference – Shaping the future: Linking policy, research and outcomes, Perth, Australia 2012
CCA (2004), Guide to Residential Streets and Paths (T51), Australia: Cement & Concrete
Association of Australia.
Chen, H., Dere, Y., Sotelino, E. and Archer, G. (2002), Mid-Panel Cracking of Portland Cement
Concrete Pavements in Indiana, Joint Transportation Research Program. Paper 146.
CSIRO & BoM (2007), Climate Change in Australia: Technical report 2007, Melbourne: CSIRO.
Department of Infrastructure and Planning (2009), South East Queensland Regional Plan 20092031, State of Queensland.
Department of Transport and Main Roads (2009), Queensland Transport and Roads Investment
Program 2010-11 to 2013-14, Queensland Government.
Friberg, B. (1938), Load and Deflection Characteristics of Dowels in Transverse Joints of
Concrete Pavements, Proceedings of Highways Research Board, Washington, DC, : National
Research Council.
Huang, B., Tutumluer, E., Al-qadi, I.L., Prozzi, J. and Shu, X. (2010), Paving Materials and
Pavement Analysis. Proceedings of sessions of GeoShanghai 2010 International Conference.
American Society of Civil Engineers Publications.
Intergovernmental Panel on Climate Change, I. (2007), Climate Change 2007: Synthesis Report
Pachauri, Cambridge, UK: Cambridge University Press.
Ju, S. (2009), Finite element investigation of traffic induced vibrations, Journal of Sound and
Vibration, 321, 837-853.
Machemehl, R., Wang, F. and Prozzi, J.A. (2005), Analytical Study of Effects of Truck Tire
Pressure on Pavements Using Measured Tire-Pavement Contact Stress Data. Proceedings for
the 84th TRB Annual Meeting. Transportation Research Board.
Serrao-Neumann, S., Low Choy, D., van Staden, R., Crick, F., Sahin, O., Guan, H. and Chai, G.
(2011), Climate Change Impacts on Road Infrastructure Systems and Services in South East
Queensland: Implications for Infrastructure Planning and Management, States of Australian
Cities (SOAC2011). University of Melbourne.
Strand7 (2004), Strand7 Theoretical Manual, Sydney, Strand7 Pty Ltd.
von Quintus, H. and Killingsworth, B. (1993), Guide for Design of Pavement Structures,
American Association of State Highway and Transportation Officials. Washington, DC.
ACKNOWLEDGMENTS
A special thank you goes to the South East Queensland Climate Adaptation Research Initiative
(SEQ-CARI) for financial support. The authors would also like to acknowledge Gold Coast City
Council, Logan City Council, Redland City Council, Queensland Department of Transport and
Main Roads-South Coast & Brisbane Metropolitan Regions and Queensland Southern Regional
Roads Group/Road Alliance for their financial contribution to the pavement research. Mr Chun
Song, an Industry Affiliate Project student at the Griffith School of Engineering is also mentioned
here for his contribution to the study.
AUTHOR BIOGRAPHIES
Dr Gary Chai
Dr Chai is the Principal Researcher of the Road Research Alliance's long-term pavement
performance study in South East Queensland. He has active research interests in the fields of
pavement design and management, pavement investigation and evaluation, finite element
© ARRB Group Ltd and Authors 2012
11
25th ARRB Conference – Shaping the future: Linking policy, research and outcomes, Perth, Australia 2012
analysis of concrete pavement and soil mechanics of pavement foundation. Dr Chai has more
than 25 years of extensive experience in road pavement design, construction and in the
management of highway projects under the Design-Build-Operate and Transfer (DBOT)
contractual scheme in the Asia Pacific Region.
Associate Professor Hong Guan
Associate Professor Guan’s main research areas are structural engineering and computational
mechanics with special interests in finite element modelling and failure analysis of concrete
structures, application of finite element method in dental implant research, bridge deterioration
modelling and structural topology optimisation. She has co-authored one edited conference
proceedings and over 120 refereed journal and conference papers.
Dr Rudi van Staden
Dr van Staden’s research interests span the fields of finite element analysis and modelling,
analysis of concrete and flexible pavement deterioration in accordance with climate change
impacts. He has established a promising publication track record in particular he has developed
research strengths in Finite Element Analysis and Linear Elastic Theory techniques and
fundamental understanding and scientific basis for pavement design and deterioration
characteristics during vehicular loading.
Professor Yew-Chaye Loo
Professor Loo's principal research interests lie in the areas of Concrete Structures, Bridge
Engineering, Computational Mechanics, Construction Materials and Design Code Development.
He has authored and co-authored four books and has signed a contract for the fifth; published
two edited conference proceedings and over 200 refereed journal and conference papers.
Copyright Licence Agreement
The Author allows ARRB Group Ltd to publish the work/s submitted for the 25th ARRB Conference,
granting ARRB the non-exclusive right to:
• publish the work in printed format
• publish the work in electronic format
• publish the work online.
The Author retains the right to use their work, illustrations (line art, photographs, figures, plates) and
research data in their own future works
The Author warrants that they are entitled to deal with the Intellectual Property Rights in the works
submitted, including clearing all third party intellectual property rights and obtaining formal permission from
their respective institutions or employers before submission, where necessary.
© ARRB Group Ltd and Authors 2012
12